Hello Kalan
This already exists and is currently being beta-tested on the toolserver. If would be very cool if you could build on top of this!
The backend consists of two components, graphcore and graphserv. WMDE contracted Johannes Kroll to implement them based on my spec. The code is available an GitHub:
* https://github.com/jkroll20/graphcore * https://github.com/jkroll20/graphserv
Documentation:
* https://github.com/jkroll20/graphcore/blob/master/README.rst * https://github.com/jkroll20/graphserv/blob/master/doc/install.rst * https://github.com/jkroll20/graphserv/blob/master/doc/usage.rst
You can see the original specs here:
* http://wikimedia.de/wiki/Ausschreibung/GraphServ/Spec
I have developed client libraries for this in Python and PHP:
* http://wikimedia.de/wiki/Ausschreibung/GraphServ/Spec
I have also started to implement a Web-API for accessing the category structure, including a component for fetching updates from the wikipedia databases:
* https://svn.toolserver.org/svnroot/daniel/duesenstuff/trunk/CatGraph/
The API is online here, but the backend is frequently unavailable or gets reset to an empty state:
* http://toolserver.org/~daniel/CatGraph/CatGraphApi.php
more info at http://toolserver.org/~daniel/CatGraph/help.html
SO, check it out, and see if you find something you can build with this! -- daniel
PS: all of this is copyrighted by wikimedia germany and available under a free license. If the licensing of any component is unclear, please let me know, i'll fix it.
On 06.04.2012 17:11, Kalan wrote:
Hello,
I present a proposal for a system that is going to solve the category intersection problem (and, in fact, allow all set-theory operations for categories). The name I suggest for it is “Caterpillar”.
Such a system would consist of a C or C++ backend daemon that stores all the category graph in RAM and MediaWiki extension frontend.
Here is the full text of the proposal: https://www.mediawiki.org/wiki/User:Kalan/gsoc
I am seeking for an interested mentor.
— Kalan